Project Manager and Entrepreneur, with experience in leading digital and organizational transformations. Currently building workflows with AI to augment and automate key processes.
I am a Project Manager with experience in building projects from idea to business case to final deployment. I am currently building with AI to simplify and automate processes for Sales & Trading stations. I am also involved in building new products using AI workflows. I currently live in Geneva, Switzerland.
Retail trading is restricted by the size of capital and the hours a trader can put into sitting in front of the desk.
With the advent of Machine Learning models and thereafter LLM's provides an opportunity to build AI based trading bots that can both trade autonomously or provide valuable information on the potential direction of the trade, saving countless hours on the desk.
The first challenge is to be able to parse through past financial data to help find the daily bias. I trained, validated and benchmarked multiple time series transformers to help with this process. It involved working with large datasets, cleaning and then building test scripts using AI coding.
The second step is then to build an agentic universe which pulls data from the time series transformers, calls relevant LLM agents to finally build a conclusion on the trade. I built this from scratch using Claude Opus' coding skills. It is a project running successfully, fully automating a traders workflow.
A food startup had a plan to run a cold calling campaign to importers in Europe. They had a list of companies that needed to be researched and contact details from public facing information needed to be fetched.
The usual way would be to hire a contractor to manually research the companies and send out emails. I helped in using n8n to build AI agents to do the research online, populate an excel sheet with details and do the first reach out.
The process that would normally have taken 15 business days was done in 6 hours at a running cost of less than $5. The AI agent compared to a human had 25% more messages opened and replied.